Usage
Comprehensive usage guidelines, including data download, processing, models description, configuration and execution, cross-validation, and verification, are available in the Training Documentation. But for a quick start, use the example notebooks.
Download example notebooks:
git clone https://github.com/hmandela/WASS2S_notebooks.git
or download the zip file:
wget https://github.com/hmandela/WASS2S_notebooks/archive/refs/heads/main.zip -O WASS2S_notebooks.zip
- Data Download & Management
- Preprocessing Modules
- Computing Predictands
- Agro-Climatic Seasonality
- Spell Analysis (Dry/Wet)
- ETCCDI Temperature Extremes
- ETCCDI Precipitation Extremes
- Input Data Formats
- Merging Gridded Data with Observations
- Prerequisites & Data Formats
- Class Initialization
- Merging Methods
- Visualization
- Usage Example
- Bias Correction Modules
- Prerequisites
- Precipitation Bias Correction (WAS_Qmap)
- Continuous Bias Correction (WAS_bias_correction)
- Data Transformation & Skewness Analysis
- Prerequisites & Input Data
- Skewness Detection and Handling
- Distribution Fitting (Clustering Approach)
- Visualization
- Models and Cross-Validation
- Linear Regression and Regularization Models
- Linear Regression
- Ridge Regression (L2 Regularization)
- Lasso Regression (L1 Regularization)
- ElasticNet (L1 + L2 Regularization)
- Probabilistic Calculation Methods
- Advanced Machine Learning Models
- Support Vector Regression (SVR)
- Multi-Layer Perceptron (MLP)
- Stacking Ensemble (RF + XGB + MLP)
- Multivariate Adaptive Regression Splines (MARS)
- Optimization Methods
- EOF Analysis & Principal Component Regression
- EOF Analysis (WAS_EOF)
- Principal Component Regression (WAS_PCR)
- CCA Models
- Analog Forecasting Methods
- Quantifying Uncertainty via Cross-Validation
- Estimating Prediction Uncertainty
- Verification Module
- Multi-Model Ensemble (MME) Techniques
- 1. Data Preparation
- 2. Weighted Ensembles (Linear)
- 3. Min et al. (2009) Probabilistic MME
- 4. Machine Learning Ensembles
- 5. Calibration & Post‑Processing
- Implementation of WAS-NextGen Approaches